Abortion study EU datasets

In this study, we will examine abortion data across Europe, with a particular focus on Romania, and aim to answer several questions:

1. How did the abortion rates change over time across EU countries?​

2. How do EU countries rank in terms of abortion numbers?​

3. How do demographics like age and region affect abortion rates?​

4. Do major events or new legislation impact the number of abortions?​

5. Is there a correlation between the natality and the number of abortions?​

​

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
from matplotlib.ticker import FuncFormatter

# Load the datasets

eu_1 = pd.read_csv('eu_ab_1.csv')
eu_2 = pd.read_csv('eu_ab_2.csv')
eu_pop = pd.read_csv('eu_population_count.csv')
eu_births = pd.read_csv('eu_live_births.csv')
eu_births_by_age = pd.read_csv('eu_births_by_age.csv')

eu_1 = eu_1[(eu_1['geo'] != 'Germany including former GDR') & (eu_1['geo'] != 'Metropolitan France')]
eu_pop = eu_pop[(eu_pop['geo'] != 'Germany including former GDR') & (eu_1['geo'] != 'Metropolitan France')]

print(eu_1.isnull().sum())
print(eu_2.isnull().sum())

DATAFLOW          0
LAST UPDATE       0
freq              0
unit              0
age               0
geo               0
TIME_PERIOD       0
OBS_VALUE         0
OBS_FLAG       7540
CONF_STATUS    7595
dtype: int64
DATAFLOW         0
LAST UPDATE      0
freq             0
indic_de         0
unit             0
geo              0
TIME_PERIOD      0
OBS_VALUE        0
OBS_FLAG       497
CONF_STATUS    502
dtype: int64

C:\Users\Tudor\AppData\Local\Temp\ipykernel_9228\79747375.py:10: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  eu_pop = eu_pop[(eu_pop['geo'] != 'Germany including former GDR') & (eu_1['geo'] != 'Metropolitan France')]

eu_1.describe()

	TIME_PERIOD	OBS_VALUE	CONF_STATUS
count	7595.000000	7.595000e+03	0.0
mean	2004.618565	1.589860e+04	NaN
std	12.723524	5.593180e+04	NaN
min	1960.000000	0.000000e+00	NaN
25%	1998.000000	1.200000e+02	NaN
50%	2006.000000	1.771000e+03	NaN
75%	2014.000000	1.046700e+04	NaN
max	2023.000000	1.407042e+06	NaN

Q1: How have abortion rates changed over time across EU countries?

abortions_by_year = eu_1[eu_1['age'] == 'Total'].groupby('TIME_PERIOD').OBS_VALUE.sum().reset_index()
abortions_by_year.columns = ['Year', 'Total Abortions']
abortions_by_year

	Year	Total Abortions
0	1960	1394005.0
1	1961	448654.0
2	1962	364640.0
3	1963	361941.0
4	1964	385124.0
...	...	...
59	2019	523981.0
60	2020	762567.0
61	2021	634463.0
62	2022	655626.0
63	2023	460367.0

64 rows × 2 columns

plt.figure(figsize=(20, 10))
plt.title('Total Abortions by Year in the EU')
plt.xlabel('Year')
plt.ylabel('Total Abortions (in milions)')
plt.plot(abortions_by_year.Year, abortions_by_year['Total Abortions']/ 1_000_000, marker='o')
plt.tight_layout()
plt.grid(True)

years = abortions_by_year['Year']
plt.xticks(np.arange(years.min(), years.max() + 2, 2))
plt.show()

It seemes that somthing might of happened right before 2010 where we had a big rise and than a fall of the abortion numbers per year
Maybe visialising every country in the eu might help us understand this better.

plt.figure(figsize=(20, 20))

index = 1
for location in eu_1.geo.unique():
    abortions_per_country_by_year = eu_1[(eu_1['geo'] == location) & (eu_1['age'] == 'Total')].groupby('TIME_PERIOD').OBS_VALUE.sum().reset_index()
    abortions_per_country_by_year.columns = ['Year', 'Total Abortions']

    plt.subplot(7, 6, index)
    index += 1
    plt.title(location)
    plt.xlabel('Year')
    plt.ylabel('Total Abortions (in milions)')
    plt.plot(abortions_per_country_by_year.Year, abortions_per_country_by_year['Total Abortions']/ 1_000_000, marker='o')
    plt.grid(True)

plt.suptitle('Total Abortions by Year in the EU', fontsize=24, y=1.02)
plt.tight_layout()
plt.show()

Not all contries seem to have a noticible raise in abortions in 2007.

But let's take only Romania:

abortions_romania = eu_1[(eu_1['geo'] == 'Romania') & (eu_1['age'] == 'Total')]

plt.figure(figsize=(20, 10))
plt.title('Total Abortions by Year in Romania')
plt.xlabel('Year')
plt.ylabel('Total Abortions (in milions)')
plt.plot(abortions_romania.TIME_PERIOD, abortions_romania['OBS_VALUE']/ 1_000_000, marker='o')
plt.tight_layout()
plt.grid(True)

years = abortions_romania['TIME_PERIOD']
plt.xticks(np.arange(years.min(), years.max() + 10, 2))
plt.show()

The decline after 1966 happened because abortion was made iligal in Romania,
While the rise in 1990 happened because it was made legal again.
After that it seems it has been on a decline -maybe- due to public awareness around sexual health, contraception, and safe sex practices..

Q2: How do EU countires rank in terms of avort numbers?

we'll only take 2023 in consideration

eu_1.groupby('TIME_PERIOD')['geo'].nunique().sort_values(ascending=False)

TIME_PERIOD
2007    28
2017    27
2006    27
2008    27
1992    26
        ..
1966     7
1961     6
1964     6
1963     6
1962     6
Name: geo, Length: 64, dtype: int64

eu_2.groupby('TIME_PERIOD')['geo'].nunique().sort_values(ascending=False).head(10)

TIME_PERIOD
2017    28
2018    27
2014    26
2015    25
2020    25
2016    24
2019    24
2022    22
2013    21
2023    20
Name: geo, dtype: int64

abortions2023 = eu_1[(eu_1['TIME_PERIOD'] == 2022) & (eu_1['age'] == 'Total')]
abortions2023 = abortions2023.groupby('geo')['OBS_VALUE'].sum().reset_index()
abortions2023.sort_values(by='OBS_VALUE', inplace=True, ascending=True)

plt.figure(figsize=(10, 6))
plt.barh(y=abortions2023.geo, width=abortions2023.OBS_VALUE)
plt.title('Abortion in the EU in 2023')
plt.xlabel('Abortion count')
plt.ylabel('Country')
plt.tight_layout()

Intresting resuls, but I think we'd get a better understanding if instead of only abortions, we'd visualise abortions per population for every country

population2023 = eu_pop[eu_pop['TIME_PERIOD'] == 2022][['geo', 'OBS_VALUE']]
population2023.sort_values(by='OBS_VALUE', inplace=True)
population2023.head(3)

	geo	OBS_VALUE
543	San Marino	33698
333	Liechtenstein	39308
5	Andorra	79535

First let's visualise the population for each country

plt.figure(figsize=(10, 8))
plt.barh(y=population2023.geo, width=population2023.OBS_VALUE/ 1_000_000)
plt.title('Population in the EU in 2023')
plt.xlabel('Population (in milions)')
plt.ylabel('Country')
plt.tight_layout()

plt.figure(figsize=(18, 8))

population2023 = population2023[(~population2023.geo.str.contains('Euro')) & (population2023.geo != 'Türkiye')]

plt.subplot(1, 2, 1)
plt.barh(abortions2023.geo, abortions2023.OBS_VALUE)
plt.title("Abortions in the EU (2023)")

plt.subplot(1, 2, 2)
plt.barh(population2023.geo, population2023.OBS_VALUE / 1_000_000)
plt.title("Population in the EU (2023)")

plt.tight_layout()
plt.show()

We can see in the two plots above that population seems to be a big factor when talking about abortion, contries with a big population will have a big number of abortions, this seems to be the trend. But there are exceptions, usually estern European cuntries such as Georgia and Moldova.
Maybe looking at the percentages will give us a better idea of what is happening.

Interesting we see that even tho Poland has a bigger Population it has more abortions

abort_per_pop2023 = pd.merge(abortions2023, population2023, on='geo', how='inner')
abort_per_pop2023['OBS_VALUE'] = (abort_per_pop2023['OBS_VALUE_x'] / abort_per_pop2023['OBS_VALUE_y']) * 1_00
abort_per_pop2023.sort_values(by='OBS_VALUE', inplace=True)
abort_per_pop2023

	geo	OBS_VALUE_x	OBS_VALUE_y	OBS_VALUE
0	Poland	160.0	36889761	0.000434
1	Albania	676.0	2793592	0.024198
2	Lithuania	2705.0	2805998	0.096401
5	Slovakia	5891.0	5434712	0.108396
13	Italy	65449.0	59030133	0.110874
8	Serbia	8090.0	6797105	0.119021
15	Germany	103927.0	83237124	0.124857
6	Finland	7782.0	5548241	0.140261
3	Latvia	2793.0	1875757	0.148900
12	Romania	28661.0	19042455	0.150511
9	Portugal	15870.0	10421117	0.152287
10	Czechia	16438.0	10516707	0.156304
14	Spain	98316.0	47486843	0.207038
4	Estonia	3386.0	1331796	0.254243
7	Moldova	7896.0	2565030	0.307833
16	France	233281.0	68091703	0.342598
11	Georgia	16619.0	3688647	0.450545

plt.figure(figsize=(12, 8))
plt.barh(y=abort_per_pop2023.geo, width=abort_per_pop2023.OBS_VALUE)
plt.title('Abortion as a percent Population in the EU in 2023')
plt.xlabel('Abortion % of pop')
plt.ylabel('Country')
plt.tight_layout()

Weird, maybe when we'll talk about natalities we will understand this graph better.

Q3: How do demographics like age and region affect abortion?

Okay now let's see how age influences abortions

abortions_by_age = eu_1[eu_1['TIME_PERIOD'] == 2022].groupby('age').OBS_VALUE.sum().reset_index()
abortions_by_age.sort_values(by='OBS_VALUE', inplace=True)
abortions_by_age

	age	OBS_VALUE
0	50 years or over	241.0
8	Less than 15 years	1589.0
10	Unknown	1953.0
7	From 45 to 49 years	4084.0
6	From 40 to 44 years	50376.0
1	From 15 to 19 years	51377.0
5	From 35 to 39 years	115562.0
2	From 20 to 24 years	129504.0
3	From 25 to 29 years	141271.0
4	From 30 to 34 years	148820.0
9	Total	655626.0

abortions_by_age.drop(9, inplace=True)  # drop toal abortion
abortions_by_age.drop(10, inplace=True) # drop Unknown 

plt.figure(figsize=(12, 8))
plt.barh(y=abortions_by_age.age, width=abortions_by_age.OBS_VALUE / 1_000)
plt.title('Abortion by age in the EU')
plt.xlabel('Abortions in thousands')
plt.ylabel('Age of women')
plt.tight_layout()

Let's take only romania

abortions_by_age_Romania2022 = eu_1[(eu_1['TIME_PERIOD'] == 2022) & (eu_1['geo'] == 'Romania')].groupby('age').OBS_VALUE.sum().reset_index()
abortions_by_age_Romania2022.sort_values(by='OBS_VALUE', inplace=True)

# abortions_by_age_Romania.drop(9, inplace=True)  # drop toal abortion
# abortions_by_age_Romania.drop(10, inplace=True) # drop Unknown 
abortions_by_age_Romania2022 = abortions_by_age_Romania2022[
    ~abortions_by_age_Romania2022['age'].isin(['Total', 'Unknown'])
]


plt.figure(figsize=(12, 8))
plt.barh(y=abortions_by_age_Romania2022.age, width=abortions_by_age_Romania2022.OBS_VALUE)
plt.title('Abortion by age in Romania in 2022')
plt.xlabel('Abortions')
plt.ylabel('Age of women')
plt.tight_layout()

Let's visualize it in a piechart

import matplotlib.cm as cm

plt.figure(figsize=(10, 8))

sizes = abortions_by_age_Romania2022.OBS_VALUE.values
labels = abortions_by_age_Romania2022.age.values

percentages = sizes / sizes.sum() * 100
def make_label(i):
    if i > 3:
        return f"{labels[i]}: {percentages[i]:.1f}%"
    else:
        return ''

pie_labels = [make_label(i) for i in range(len(labels))]

def func(pct):
    return f'{pct:.1f}%' if pct > 5 else ''

num_slices = len(abortions_by_age_Romania2022) + 1
cmap = cm.get_cmap('viridis')
colors = cmap(np.linspace(0, 1, num_slices))

wedges, texts, autotexts = plt.pie(
    abortions_by_age_Romania2022.OBS_VALUE,
    autopct=func,
    colors=colors,
    startangle=140,
    wedgeprops={'edgecolor': 'black'}
)

sorted_indices = np.argsort(-sizes)
legend_labels = [f'{labels[i]}: {percentages[i]:.1f}%' for i in sorted_indices]
sorted_wedges = [wedges[i] for i in sorted_indices]

plt.legend(sorted_wedges, legend_labels, title="Age Groups", loc="center left", bbox_to_anchor=(1, 0, 0.5, 1))

plt.title('Abortion distribution by age group in Romania (2022)')
plt.tight_layout()
plt.show()

C:\Users\Tudor\AppData\Local\Temp\ipykernel_9228\4029927240.py:21: MatplotlibDeprecationWarning: The get_cmap function was deprecated in Matplotlib 3.7 and will be removed in 3.11. Use ``matplotlib.colormaps[name]`` or ``matplotlib.colormaps.get_cmap()`` or ``pyplot.get_cmap()`` instead.
  cmap = cm.get_cmap('viridis')

First question that I have, did this change over the time?

plt.figure(figsize=(20, 40))

index = 1
years = eu_1.TIME_PERIOD.unique()
years.sort()
for year in years:

    abortions_by_age_Romania = eu_1[(eu_1['TIME_PERIOD'] == year) & (eu_1['geo'] == 'Romania')].groupby('age').OBS_VALUE.sum().reset_index()
    abortions_by_age_Romania.sort_values(by='OBS_VALUE', inplace=True)

    abortions_by_age_Romania = abortions_by_age_Romania[
        ~abortions_by_age_Romania['age'].isin(['Total', 'Unknown'])
    ]

    if abortions_by_age_Romania.empty:
        continue

    plt.subplot(20, 3, index)
    index += 1

    plt.barh(y=abortions_by_age_Romania.age, width=abortions_by_age_Romania.OBS_VALUE)
    plt.title(f'Abortion by age in Romania in {year}')
    plt.xlabel('Abortions')
    plt.ylabel('Age of women')
    plt.tight_layout()

plt.suptitle('Abortions ages each Year in Romania', fontsize=24, y=1.02)
plt.tight_layout()
plt.show()

Alright this does not tell us a lot let's use a heatmap

age_order = [
    'Less than 15 years', 'From 15 to 19 years', 'From 20 to 24 years', 'From 25 to 29 years', 'From 30 to 34 years',
    'From 35 to 39 years', 'From 40 to 44 years', 'From 45 to 49 years', '50 years and over'
]

import seaborn as sns
import matplotlib.pyplot as plt

romania_ab = eu_1[eu_1['geo'] == 'Romania'].copy()
romania_ab = romania_ab[~romania_ab['age'].isin(['Total', 'Unknown'])]
heatmap_data = romania_ab.pivot_table(index='age', columns='TIME_PERIOD', values='OBS_VALUE', aggfunc='sum')
heatmap_data = heatmap_data.sort_index(axis=1)
heatmap_data = heatmap_data.reindex(age_order[::-1])

plt.figure(figsize=(14, 8))
sns.heatmap(heatmap_data, cmap='Reds', linewidths=0.5)
plt.title('Abortions by Age Group and Year in Romania', fontsize=20)
plt.xlabel('Year')
plt.ylabel('Age Group')
plt.tight_layout()
plt.show()

time_restriction = [1995, 2016]
index = 1
plt.figure(figsize=(25, 10))

for first_year in time_restriction:
    plt.subplot(1, 2, index)
    index += 1

    heatmap_data = romania_ab[romania_ab['TIME_PERIOD'] >= first_year].pivot_table(
        index='TIME_PERIOD', columns='age', values='OBS_VALUE', aggfunc='sum'
    )

    heatmap_data.index = heatmap_data.index.astype(int)
    heatmap_data = heatmap_data.sort_index()
    heatmap_data = heatmap_data.reindex(columns=age_order, fill_value=0)
    heatmap_data = heatmap_data[age_order]

    average_abortions = heatmap_data.mean().sort_values(ascending=False)
    sorted_age_order = average_abortions.index.tolist()

    for age_group in sorted_age_order:
        plt.plot(heatmap_data.index, heatmap_data[age_group] / 1_000, label=age_group, marker='o')

    plt.title(f'From {first_year} to 2023', fontsize=20)
    plt.xlabel('Year')
    plt.ylabel('Number of Abortions in thousands')
    plt.legend(title='Age Group')
    plt.grid(True)
    
plt.suptitle('Trends in Abortion Numbers by Age Group in Romania (Yearly)', fontsize=16, y=1)
plt.tight_layout() 
plt.show()

We can see a decreasing trend in the number of abortions in Romania, by most if not all age groups.

Q4: Do major events or new legislation impact the number of abortions?

1.Romania

Abortion was legalized in Romania in 1957. However, the situation was complex, with the legalization being followed by a ban on abortion in 1966, according to an article in balcanicaucaso.org. The 1966 ban, Decree 770, was put in place to counter a decline in the birth rate. The ban was later lifted in 1989 after the fall of the communist regime.

plt.figure(figsize=(20, 10))
plt.title('Total Abortions by Year in Romania')
plt.xlabel('Year')
plt.ylabel('Total Abortions (in milions)')
plt.plot(abortions_romania.TIME_PERIOD, abortions_romania['OBS_VALUE']/ 1_000_000, marker='o')
plt.tight_layout()
plt.grid(True)

plt.axvline(x = 1957, color = 'orange', label='abortion made legal in Romania')
plt.axvline(x = 1966, color = 'red', label='abortion made iligal in Romania')
plt.axvline(x = 1989, color = 'green', label='abortion made legal again in Romania')

years = abortions_romania['TIME_PERIOD']
plt.xticks(np.arange(years.min() - 5, years.max() + 10, 2))
plt.legend()
plt.show()

Observations:

1. Although data prior to 1957 is unavailable, we observe a rising trend in abortion rates from 1960 to 1965 — the year before abortion was criminalized.

2. In 1966, abortion was made illegal in Romania, leading to a sharp and sustained decline in recorded abortion rates. From that point onward, abortions were permitted only under strict medical exceptions. As a result, many women turned to unsafe, illegal procedures, and the ban is estimated to have caused the deaths of approximately 10,000 women. Due to the underground nature of these procedures, it is impossible to determine the exact number of illegal abortions that occurred between 1966 and 1989.

3. Following the fall of the Communist regime in 1989, abortion was legalized again. This led to a big increase in recorded abortion rates — not only because access was restored, but also because previously undocumented procedures could now be officially registered.

2.Poland

Abortion was first legalised in Poland in 1956 and after that it was made iligal in 1993, now let's see poland and also it's neibors arbortion number by years:

import numpy as np
import matplotlib.pyplot as plt

abortions_pl = eu_1[(eu_1['geo'] == 'Poland') & (eu_1['age'] == 'Total')]
abortions_gr = eu_1[(eu_1['geo'] == 'Germany') & (eu_1['age'] == 'Total')]
abortions_lt = eu_1[(eu_1['geo'] == 'Latvia') & (eu_1['age'] == 'Total')]
abortions_cz = eu_1[(eu_1['geo'] == 'Czechia') & (eu_1['age'] == 'Total')]
abortions_sl = eu_1[(eu_1['geo'] == 'Slovakia') & (eu_1['age'] == 'Total')]
abortions_bl = eu_1[(eu_1['geo'] == 'Belarus') & (eu_1['age'] == 'Total')]

plt.figure(figsize=(10, 5))
plt.title('Total Abortions by Year in Poland and Its Neighbors', fontsize=20)
plt.xlabel('Year')
plt.ylabel('Total Abortions (in thousands)')

plt.plot(abortions_pl.TIME_PERIOD, abortions_pl['OBS_VALUE'] / 1_000, marker='o', label='Poland')
plt.plot(abortions_gr.TIME_PERIOD, abortions_gr['OBS_VALUE'] / 1_000, marker='o', label='Germany')
plt.plot(abortions_lt.TIME_PERIOD, abortions_lt['OBS_VALUE'] / 1_000, marker='o', label='Latvia')
plt.plot(abortions_cz.TIME_PERIOD, abortions_cz['OBS_VALUE'] / 1_000, marker='o', label='Czechia')
plt.plot(abortions_sl.TIME_PERIOD, abortions_sl['OBS_VALUE'] / 1_000, marker='o', label='Slovakia')
plt.plot(abortions_bl.TIME_PERIOD, abortions_bl['OBS_VALUE'] / 1_000, marker='o', label='Belarus')


plt.axvline(x = 1956, color = 'gray', label='abortion made ligal in Poland')
plt.axvline(x = 1993, color = 'black', label='abortion made iligal in Poland')

plt.tight_layout()
plt.grid(True)
plt.legend(loc='center left', bbox_to_anchor=(1, 0.5))
years = abortions_pl['TIME_PERIOD']
plt.xticks(np.arange(years.min(), years.max() + 1, 5))  # minor fix: +1 not +10
plt.show()

Alright so it does not seem to be any semnificative rise in the abortion rates of its neighbors when the law was passed

My previous observation might be wrong, the anti abortion agenda in Poland gained power in 1989 after the fall of Comunism in Poland
In 1989 after Comunism fell, there were a lot of anti abortion laws implemented
This movment started in Late 1970s – Early 1980, with a Solidarity Movement beeing started in 1980 by the the Catholic Church

import numpy as np
import matplotlib.pyplot as plt

abortions_pl = eu_1[(eu_1['geo'] == 'Poland') & (eu_1['age'] == 'Total')]
abortions_gr = eu_1[(eu_1['geo'] == 'Germany') & (eu_1['age'] == 'Total')]
#abortions_lt = eu_1[(eu_1['geo'] == 'Latvia') & (eu_1['age'] == 'Total')]
abortions_cz = eu_1[(eu_1['geo'] == 'Czechia') & (eu_1['age'] == 'Total')]
abortions_sl = eu_1[(eu_1['geo'] == 'Slovakia') & (eu_1['age'] == 'Total')]
#abortions_bl = eu_1[(eu_1['geo'] == 'Belarus') & (eu_1['age'] == 'Total')]

plt.figure(figsize=(10, 5))
plt.title('Total Abortions by Year in Poland and Its Neighbors', fontsize=20)
plt.xlabel('Year')
plt.ylabel('Total Abortions (in thousands)')

plt.plot(abortions_pl.TIME_PERIOD, abortions_pl['OBS_VALUE'] / 1_000, marker='o', label='Poland')
plt.plot(abortions_gr.TIME_PERIOD, abortions_gr['OBS_VALUE'] / 1_000, marker='o', label='Germany')
#plt.plot(abortions_lt.TIME_PERIOD, abortions_lt['OBS_VALUE'] / 1_000, marker='o', label='Latvia')
plt.plot(abortions_cz.TIME_PERIOD, abortions_cz['OBS_VALUE'] / 1_000, marker='o', label='Czechia')
plt.plot(abortions_sl.TIME_PERIOD, abortions_sl['OBS_VALUE'] / 1_000, marker='o', label='Slovakia')
#plt.plot(abortions_bl.TIME_PERIOD, abortions_bl['OBS_VALUE'] / 1_000, marker='o', label='Belarus')


plt.axvline(x = 1980, color = 'gray', label='anti abortion movement')
plt.axvline(x = 1989, color = 'dimgray', label='anti abortion laws implemented')
plt.axvline(x = 1993, color = 'black', label='Abortion made Iligal')

plt.tight_layout()
plt.grid(True)
plt.legend()
years = abortions_pl['TIME_PERIOD']
plt.xticks(np.arange(years.min(), years.max() + 1, 5))  # minor fix: +1 not +10
plt.show()

My intuition made me think, that people that can't do abortion in Poland anymore will go to the neighboring countries
But we have to take into account that Poland was not part of EU untill 2004 so to leave the country to get an abortion was very hard at that point.

Q5: Is there a corelation between the natality and the number of abortions?

eu_births.head()

	DATAFLOW	LAST UPDATE	freq	indic_de	geo	TIME_PERIOD	OBS_VALUE	OBS_FLAG	CONF_STATUS
0	ESTAT:EQ_FER05(1.0)	27/03/25 23:00:00	Annual	Live births - females	Andorra	2018	250	NaN	NaN
1	ESTAT:EQ_FER05(1.0)	27/03/25 23:00:00	Annual	Live births - females	Albania	2013	17089	NaN	NaN
2	ESTAT:EQ_FER05(1.0)	27/03/25 23:00:00	Annual	Live births - females	Albania	2014	17077	NaN	NaN
3	ESTAT:EQ_FER05(1.0)	27/03/25 23:00:00	Annual	Live births - females	Albania	2015	15612	NaN	NaN
4	ESTAT:EQ_FER05(1.0)	27/03/25 23:00:00	Annual	Live births - females	Albania	2016	15332	NaN	NaN

Okay let's first take a look at what is the treand of birthrate in Europe

euro_area = eu_births[eu_births['geo'] == 'Euro area – 20 countries (from 2023)']

plt.figure(figsize=(20, 10))
plt.title('Total Births by Year in Euro area')
plt.xlabel('Year')
plt.ylabel('Total Births (in milions)')
plt.plot(euro_area.TIME_PERIOD, euro_area['OBS_VALUE']/ 1_000_000, marker='o')
plt.tight_layout()
plt.grid(True)

years = euro_area['TIME_PERIOD']
plt.xticks(np.arange(years.min(), years.max() + 1, 2))
plt.show()

It seems Births are on a decline in europe, nothing new here

Let's see how each contre is doing:

plt.figure(figsize=(20, 20))

index = 1
for location in eu_births.geo.unique():
    births_per_country_by_year = eu_births[eu_births['geo'] == location].groupby('TIME_PERIOD').OBS_VALUE.sum().reset_index()
    births_per_country_by_year.columns = ['Year', 'Total Births']

    plt.subplot(10, 5, index)
    index += 1
    plt.title(location)
    plt.xlabel('Year')
    plt.ylabel('Total Births (in thousands)')
    plt.plot(births_per_country_by_year.Year, births_per_country_by_year['Total Births']/ 1_000, marker='o')
    plt.grid(True)

plt.suptitle('Total Births by Year in the EU', fontsize=24, y=1.02)
plt.tight_layout()
plt.show()

The decline in birth rates seems to be a consistent pattern throughout Europe.

Let's take a look at Romania

romania_births = eu_births[eu_births['geo'] == 'Romania']

plt.figure(figsize=(10, 5))
plt.title('Total Births by Year in Romania')
plt.xlabel('Year')
plt.ylabel('Total Births (in thousands)')
plt.plot(romania_births.TIME_PERIOD, romania_births['OBS_VALUE']/ 1_000, marker='o')
plt.tight_layout()
plt.grid(True)

years = romania_births['TIME_PERIOD']
plt.xticks(np.arange(years.min(), years.max() + 1, 2))
plt.show()

Let's put it side by side with the abortion (starting in 2013 since the data for births starts in 2013)

romania_births = eu_births[eu_births['geo'] == 'Romania']

plt.figure(figsize=(10, 5))


plt.subplot(1, 2, 1)
plt.title('Total Births by Year in Romania')
plt.xlabel('Year')
plt.ylabel('Total Births (in thousands)')
plt.plot(romania_births.TIME_PERIOD, romania_births['OBS_VALUE']/ 1_000, marker='o')
plt.grid(True)

years = romania_births['TIME_PERIOD']
plt.xticks(np.arange(years.min(), years.max() + 1, 2))

abortions_romania_after2013 = abortions_romania[abortions_romania['TIME_PERIOD'] >= 2013]
plt.subplot(1, 2, 2)
plt.title('Total Abortions by Year in Romania')
plt.xlabel('Year')
plt.ylabel('Total Abortions (in thousands)')
plt.plot(abortions_romania_after2013.TIME_PERIOD, abortions_romania_after2013['OBS_VALUE']/ 1_000, marker='o')
plt.grid(True)

plt.xticks(np.arange(years.min(), years.max() + 1, 2))
plt.tight_layout()
plt.show()

We will analyze the share of abortions as a percentage of total pregnancy outcomes by year in Romania:

pregnancy_data = pd.merge(
    abortions_romania_after2013[['TIME_PERIOD', 'OBS_VALUE']],
    romania_births[['TIME_PERIOD', 'OBS_VALUE']],
    on='TIME_PERIOD',
    suffixes=('_abortions', '_births')
)

pregnancy_data['abortion_percentage'] = (
    pregnancy_data['OBS_VALUE_abortions'] /
    (pregnancy_data['OBS_VALUE_abortions'] + pregnancy_data['OBS_VALUE_births'])
) * 100

plt.figure(figsize=(10, 5))
plt.title("Abortion as a procentage of 'known pregnancy outcomes'")
plt.xlabel('Year')
plt.ylabel('Procentage of Abortions %')
plt.plot(pregnancy_data.TIME_PERIOD, pregnancy_data['abortion_percentage'], marker='o')
plt.tight_layout()
plt.grid(True)

years = pregnancy_data['TIME_PERIOD']
plt.xticks(np.arange(years.min(), years.max() + 1, 2))
plt.show()

I want to clarify here: 'known pregnancy outcomes' means to the sum of abortion and birth, it does not acount for miscarriages, stillbirths, and other pregnancy outcomes

We can se that this procentage was on a decline until 2021

plt.figure(figsize=(18, 8))

eu_births2023 = eu_births[(~eu_births.geo.str.contains('Euro')) & (eu_births.geo != 'Türkiye') & (eu_births.TIME_PERIOD == 2023)]
eu_births2023.sort_values(by='OBS_VALUE', inplace=True)

plt.subplot(1, 2, 1)
plt.barh(abortions2023.geo, abortions2023.OBS_VALUE)
plt.title("Abortions in the EU (2023)")

plt.subplot(1, 2, 2)
plt.barh(eu_births2023.geo, eu_births2023.OBS_VALUE / 1_000_000)
plt.title("Births in the EU (2023)")

plt.tight_layout()
plt.show()

C:\Users\Tudor\AppData\Local\Temp\ipykernel_9228\3993873374.py:4: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  eu_births2023.sort_values(by='OBS_VALUE', inplace=True)